The Bulgarian National Bank’s experience with loan- by-loan data collection for the compilation of external statistics Emil Dimitrov1 Abstract The purpose of this paper and the related presentation is to prove that microdata are very important for statistics. However, statistical data (including that based on microdata) needs first a common methodology, so as to be compiled with the expected reliability and quality and to be comparable among countries. 1 Statistics Department, Bulgarian National Bank,
[email protected]. IFC Bulletin No 37 103 Introduction It is a general rule that reliable statistics are a cornerstone for evidence-based policy making. In a period of global financial crises and increased uncertainty, the general public is confronted with a loss of confidence in the economy, and the users look more carefully at the published data. Data quality is especially important in areas linked to the package on economic governance (the so-called “Six-pack”), particularly the indicators in the Macroeconomic Imbalances Procedure (MIP) Scoreboard. Under the Alert Mechanism, the quality of data becomes very important at both national and European levels: users at the national level are very much interested in knowing whether data among countries are statistically comparable; users at the European level want scoreboard data among countries to be comparable, as this increases the reliability of the Alert Mechanism Report (AMR) and the whole MIP mechanism. Moreover, international investors and rating agencies are willing to use AMR’s reports for their analysis if they consider statistical data sufficiently reliable. The European Statistics Code of Practice and data quality The European Statistics Code of Practice (CoP) is based on 15 Principles covering the institutional environment, the statistical production processes and the output of statistics.